5 local total_attempts=$1; shift
8 echo "[NOTE] Waiting for cmd to return success: ${cmdstr}"
9 # shellcheck disable=SC2034
10 for attempt in $(seq "${total_attempts}"); do
11 # shellcheck disable=SC2015
12 eval "${cmdstr}" && break || true
13 echo -n '.'; sleep "${sleep_time}"
17 # Wait for MaaS commissioning/deploy to finish, retry on failure
18 function maas_fixup() {
19 local statuscmd="salt 'mas01*' --out yaml state.apply maas.machines.status"
20 # shellcheck disable=SC2155
21 local ncount=$(salt --out yaml 'mas01*' pillar.get maas:region:machines | \
22 grep -cE '^\s{2}\w+:$')
23 wait_for 180 "${statuscmd} | tee /dev/stderr | " \
24 "grep -Eq '((Deployed|Ready): ${ncount}|status:Failed|status:Allocated)'"
25 # shellcheck disable=SC2155
26 local statusout=$(eval "${statuscmd}")
28 # shellcheck disable=SC2155
29 local fcnodes=$(echo "${statusout}" | \
30 grep -Po '(?<=system_id:)(.*)(?=,status:Failed commissioning)')
31 for node_system_id in ${fcnodes}; do
32 salt -C 'mas01*' state.apply maas.machines.delete \
33 pillar="{'system_id': '${node_system_id}'}"
36 if [ -n "${fcnodes}" ]; then
37 salt -C 'mas01*' state.apply maas.machines
41 # shellcheck disable=SC2155
42 local fdnodes=$(echo "${statusout}" | \
43 grep -Po '(?<=system_id:)(.*)(?=,status:(Failed deployment|Allocated))')
44 for node_system_id in ${fdnodes}; do
45 salt -C 'mas01*' state.apply maas.machines.mark_broken_fixed \
46 pillar="{'system_id': '${node_system_id}'}"
49 if [ -n "${fdnodes}" ]; then
50 salt -C 'mas01*' state.apply maas.machines.deploy
57 # MaaS rack/region controller, node commissioning
58 salt -C 'mas01*' cmd.run "add-apt-repository ppa:maas/stable"
60 salt -C 'mas01*' state.apply linux,salt,openssh,ntp
61 salt -C 'mas01*' state.apply linux.network.interface
62 salt -C 'mas01*' state.apply maas.pxe_nat
63 salt -C 'mas01*' state.apply maas.cluster
64 salt -C 'cfg01*' state.apply maas.pxe_route
66 wait_for 10 "salt -C 'mas01*' state.apply maas.region"
68 salt -C 'mas01*' state.apply maas.machines
69 wait_for 10 maas_fixup
71 # cleanup outdated salt keys
72 salt-key --out yaml | awk '!/^(minions|- cfg01|- mas01)/ {print $2}' | \
73 xargs -I{} salt-key -yd {}
75 # MaaS node deployment
76 salt -C 'mas01*' state.apply maas.machines.deploy
77 wait_for 10 maas_fixup
79 salt -C 'mas01*' pillar.item\
80 maas:region:admin:username \
81 maas:region:admin:password
83 # KVM, compute node prereqs (libvirt first), VCP deployment
84 salt -C '* and not cfg01* and not mas01*' saltutil.sync_all
86 salt -C 'kvm*' pkg.install bridge-utils
87 salt -C 'kvm*' state.apply linux.network
88 salt -C 'kvm*' system.reboot
89 wait_for 90 "! salt 'kvm*' test.ping | tee /dev/stderr | fgrep -q 'Not connected'"
91 salt -C '* and not cfg01* and not mas01*' state.apply linux,ntp
93 salt -C 'kvm*' state.sls libvirt
95 salt -C '* and not cfg01* and not mas01*' state.apply salt
96 salt -C 'kvm*' saltutil.sync_all
97 wait_for 10 "! salt -C 'kvm*' state.sls salt.control | " \
98 "tee /dev/stderr | fgrep -q 'Not connected'"
100 vcp_nodes=$(salt --out yaml 'kvm01*' pillar.get salt:control:cluster:internal:node | \
101 awk '/\s+\w+:$/ {gsub(/:$/, "*"); print $1}')
103 # Check all vcp nodes are available
105 while [ $rc -ne 0 ]; do
107 for node in $vcp_nodes; do
108 salt "$node" test.ping 2>/dev/null || { rc=$?; break; };
113 wait_for 10 "salt -C '* and not cfg01* and not mas01*' saltutil.sync_all"
114 wait_for 10 "salt -C 'E@^(?!cfg01|mas01|kvm|cmp00).*' state.apply salt"
115 wait_for 10 "! salt -C 'E@^(?!cfg01|mas01|kvm|cmp00).*' state.apply linux,ntp | " \
116 "tee /dev/stderr | fgrep -q 'Not connected'"
118 wait_for 10 "salt -C 'E@^(?!cfg01|mas01|kvm|cmp00).*' ssh.set_auth_key ${SUDO_USER} \
119 $(awk 'NR==1{print $2}' "$(eval echo "~${SUDO_USER}/.ssh/authorized_keys")")"